摘要
油气长输管道环焊缝一直以来是管道的薄弱环节,容易发生失效。根据环焊缝开挖数据,对大量环焊缝相关数据进行了预处理,并采用机器学习中的随机森林方法对主要因素进行了分析,建立了环焊缝质量不合格性预测模型,并通过不同算法模型进行了验证。结果表明,构建的模型可以实现环焊缝质量不合格预测。对环焊缝的数据分析和机器学习模型的构建,可以提高环焊缝开挖准确率,节省开挖费用,为后续开挖和修复工作提供技术支持。
Girth weld of long-distance oil and gas pipeline has always been the weak link of the pipeline and is easy to fail.According to the girth weld excavation data,this paper preprocesses a large number of girth weld related data,analyzes the main factors by using random forest algorithm,constructs the unqualified girth weld quality prediction model,and verifies the model through different algorithm models.The results show that the model can predict the unqualified girth weld quality.Through the data analysis of girth weld and the construction of machine learning model,the accuracy of girth weld excavation can be improved,the cost of excavation can be saved,and the technical support can be provided for the subsequent excavation and repair work.
作者
刘亮
李娟
贺建
杨新超
吴张中
戴联双
李海润
孟祥海
LIU Liang;LI Juan;HE Jian;YANG Xinchao;WU Zhangzhong;DAI Lianshuang;LI Hairun;MENG Xianghai(PipeChina Engineering Technology Innovation Co.Ltd.,Tianjin 300450,China;Nanjing University Department of Computer Science and Technology,Nanjing,Jiangsu 210033,China;PipeChina Research Institute,Langfang,Hebei 065000,China;PipeChina Production Department,Beijing 100097,China)
出处
《石油管材与仪器》
2023年第5期64-67,73,共5页
Petroleum Tubular Goods & Instruments
基金
国家管网北方管道有限责任公司科技研究项目“基于内检测的管道环焊缝综合评价与修复技术研究”(编号:20190301)。
关键词
环焊缝
随机森林算法
机器学习
数据资产
智能管道
girth weld
random forest algorithm
machine learning
data asset
intelligent pipeline